1 INTRODUCTION

Dosidicus gigas is one of the most important target commercial species in the East Pacific (Nigmatullin et al., 2001; Arkhipkin et al., 2015b). As a predator and prey, D. gigas plays an important role in marine ecosystems (Lu and Ickeringill, 2002; Field et al., 2007).

It is the largest and one of the most abundant ommastrephid squid and widely distributed in the East Pacific from the Gulf of Alaska to southern Chile (Taipe et al., 2001; Field et al., 2007; Zeidberg and Robison, 2007). Its distribution may expand during warm and cold periods combined with the decline of predators throughout the Pacific (Zeidberg and Robison, 2007; Keyl et al., 2008). In addition to the distribution, the adult size and life span of D. gigas can be influenced by the sea temperature (Hoving et al., 2013). The life span of D. gigas is about 1 year, but some huge specimens of the large size group can live up to 1.5 to 2 years (Nigmatullin et al., 2001; Arkhipkin et al., 2015a). Significant negative effect of sea surface temperatures on the life span could lead to a shorter life cycle of squid with a smaller size (higher ambient temperature) or a longer life cycle with a larger size (lower ambient temperature) (Arkhipkin et al., 2015a). The population structure of D. gigas tends to be complex, with three intraspecific groups that can be distinguished based on the adult sizes (Nigmatullin et al., 2001). In addition, the population of D. gigas can be divided into southern and northern groups based on an analysis of randomly amplified polymorphic DNA (RAPD) and mitochondrial DNA (Sandoval-Castellanos et al., 2007, 2010).

Beaks, mainly composed of proteins and chitin fibers, are the important feeding organ of squids (Clarke, 1962; Wolff, 1984; Miserez et al., 2007). They are hard tissues with constant shape and high stiffness and can record information on life history (Guerra et al., 2010; Ikica et al., 2014). The pigmented part of the beak was harder than the transparent part based on tensile and nanoindentation tests, which might result from different contents of catechols and water in the beak (Miserez et al., 2008). In previous studies, pigmentation stages were defined based on the development of pigmentation of beaks (HernańdezGarcía et al., 1998; García, 2003; Fang et al., 2016). The darkening process was suggested to make beaks harder and stronger, which can affect feeding habits and behavior of squid (Castro and Hernández-García, 1995; García, 2003; Franco-Santos and Vidal, 2014). Regular growth lines were found in the rostrum sagittal section and inner surface of the lateral wall, and growth increment was found to be deposited daily, providing an opportunity to estimate the age of squid using beaks (Rodríguez-Domínguez et al., 2013; Villegas Bárcenas et al., 2014; Hu et al., 2016b). Furthermore, the morphology of beaks was widely used for interspecies and intraspecies identification due to their stable configuration (Martínez et al., 2002; Liu et al., 2015; Fang et al., 2017), and body size of squids can be estimated by beaks remains in the stomachs of predators (Wolff, 1984). In previous studies, the beak growth was analyzed for many species, suggesting that the development of beaks might reflect the shifts of the size and types of prey (Jackson and McKinnon, 1996; Lefkaditou and Bekas, 2004; Hu et al., 2016a).

Dosidicus gigas is a short-lived species with its abundance and distribution being easily influenced by the environmental variability (Waluda and Rodhouse, 2006; Yu et al., 2017). Waluda and Rodhouse (2006) suggested that the abundance of D. gigas was strongly influenced by mesoscale variability associated with the ENSO, and the low catches were observed during strong El Niño events of 1997 to 1998 off Peru. Robinson et al. (2013) found that the catches were high in a La Niña year (1999) associated with cold sea surface temperature (SST) and high chlorophyll a (Chl a) concentrations. Yu et al. (2016) studied the effects of the climate variability on the habitat suitability of D. gigas and found that an El Niño event weakened upwelling coupled with warm and nutrient-depleted waters, which was unfavorable for squid habitat, leading to reduced catches. However, a La Niña event strengthened upwelling coupled with cool and nutrient-enhanced waters, which was favorable for squid habitat, yielding high catches. In previous studies, the effects of environment variables on the growth of squids were studied (Pecl et al., 2004). Argüelles and Tafur (2010) analyzed the somatic and reproductive investment of D. gigas in Peruvian waters and found that the somatic and reproductive investment varied both among different years and between genders. In the Gulf of California and Mexican waters, the somatic condition of D. gigas among different years was studied using the relative weight (Wr) and Fulton's condition coefficient (K), and the K value was found more sensitive and effective in the estimation of the somatic condition of D. gigas (Bazzino Ferreri, 2014). The poor somatic condition was observed during 1998 and 1999 with unfavorable environmental conditions caused by El Niño event during 1997–1998, and the condition factors differed between genders. The sexual differences in growth for squids were commonly observed in previous studies (Markaida et al., 2004; Liu et al., 2013; Fang et al., 2017). Compared with males, females were found to have a longer life span with larger body size (Mejia-Rebollo et al., 2008).

The objectives of this study were 1) to analyze the inter-annual variability of beak morphology and the link with climate variability on the fishing ground, and 2) to explore the response mechanism of the somatic condition and abundance of D. gigas to an El Niño event.

2 MATERIAL AND METHOD 2.1 Sampling

A total of 1 096 Jumbo squid were randomly sampled from the catch of Chinese commercial jigging vessels off the Peruvian Exclusive Economic Zone (EEZ) in 2013, 2014, and 2016 (Table 1, Fig. 1). The sampled squids were immediately frozen aboard the vessels. After defrosting the samples in the laboratory, the mantle length (ML) and body weight (BW) were measured to the nearest 1 mm and 1 g, respectively. Meanwhile, sex was identified based on the approach defined by Lipiński and Underhill (1995).

2.2 Relationships between beak variables and mantle length

Both the upper and lower beaks were extracted from the buccal mass and stored in 70% alcohol. Twelve morphometric measurements were taken for each beak using digital calipers (Fig. 2), and six beak variables (UHL, UCL, ULWL, LCL, LRL, and LLWL, as seen in Fig. 2) were chosen in this study because they can be measured precisely using calipers (Fang et al., 2015).

The linear, power, exponential, and logarithmic function were used to develop the relationships between ML and beak variables for D. gigas, and the linear regression model was selected as the best model based on the value of Akaike's information criterion (AIC) (Hu et al., 2017). In this study, the relationships between ML and beak variables for squids in three years were performed using linear regression models (Hu et al., 2017):

where Y is one of the beak variables, ML is the mantle length of squid. a and b are the parameters to be estimated representing intercept and slope, respectively. ε is the error of the linear model. After simulating the linear models, an analysis of covariance (ANCOVA) was used to test the differences of slopes for different genders and years.

2.3 Somatic condition

The Fulton's condition coefficient (K) was conducted to evaluate the somatic condition of D. gigas. The K was determined from the following formula (Froese, 2006):

where K is Fulton's condition coefficient, W is body weight (g), and L is mantle length (cm). The differences of the somatic condition of D. gigas for different genders and years were analyzed using t-test (two-tailed test).

2.4 Fishery data

The D. gigas fishery data from the fishing area were obtained from the Chinese Squid-Jigging Technology Group of Shanghai Ocean University. Data included a catch (tons), fishing effort (in fishing days), and fishing location (latitude and longitude). The nominal catch per unit effort (CPUE) was calculated using the following equation (Chen et al., 2010):

where CPUE_yij is the yearly nominal CPUE (tons (t) / days (d)) at longitude i, latitude j in year y; ΣCatch_yij the total catch for all the fishing vessels within a fishing grid at longitude i, latitude j in year y; ΣEffort_yij is the sum of fishing days for all the fishing vessels within a fishing grid at longitude i, latitude j in year y.

2.5 Climate index and Chl a data

The definition for an El Niño event was based on the 3-month running mean of SST anomalies in the Niño 3.4 region (5°N–5°S, 120°–170°W) (Yu et al., 2016). An El Niño event was determined by the SSTs above a threshold of +0.5 over at least 5 consecutive months. The Oceanic Niño 3.4 indices from 2013 to 2016 were achieved from the NOAA Climate Prediction Center (http://origin.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ONI_v5.php). The monthly average Chl a data on the fishing area from 2013 to 2016 were derived from the National Oceanic and Atmospheric Administration (NOAA) Ocean Watch (http://oceanwatch.pifsc.noaa.gov/). The spatial resolution was 0.05°×0.05°.

3 RESULT 3.1 Slopes of beak variables in relation to ML

The parameters in the linear regression models between ML and beak variables were estimated for females and males sampled during 2013, 2014, and 2016 (Table 2), and the slopes for such relationships were compared between years using an analysis of covariance (ANCOVA) (Fig. 3).

In this study, the variation of slopes for beak variables in relation to ML among years was observed, while there was no significant difference (Fig. 3, ANCOVA, P > 0.05). In 2013, slopes of UCL and LRL for females were significantly greater than those of males (ANCOVA, P < 0.05), other beak variables had no significant difference between females and males (ANCOVA, P > 0.05). In 2014 and 2016, all beak variables' slopes for females were greater than those of males, while no significant difference was found (ANCOVA, P > 0.05).

3.2 The variability in somatic condition

The average K values of females and males from 2013 were 2.80±0.24 and 2.69±0.18, respectively (Table 3). The average K values of females and males caught during 2014 were 2.79±0.31 and 2.74±0.32, respectively (Table 3). The average K values of females and males sampled during 2016 were 2.57±0.23 and 2.55±0.25, respectively (Table 3).

K values of females and males during 2013 and 2014 were significantly greater than those of 2016, respectively (t-test, P < 0.01) (Fig. 4), there was no significant difference between 2013 and 2014 (t-test, P >0.05) (Fig. 4). In 2013, K values of females were significantly greater than those of males (ANCOVA, P < 0.05). In 2014 and 2016, K values of females were greater than those of males, while no significant difference was found between females and males (ANCOVA, P > 0.05).

3.3 The variability of Chl a concentration linked with El Niño event

In the normal years (i.e., 2013 and 2014), monthly average Chl a concentration showed an N-shaped variability from January to December on the fishing area (Fig. 5), increasing in the first three months followed with a decrease from March to July and an increase from July to December (Fig. 5).

In the El Niño years (i.e., 2015 and 2016), the Niño 3.4 indices were slightly greater than 0.5 in the first three months of 2015, and the Chl a concentration fluctuated (Fig. 5). Then the Niño 3.4 index increased rapidly reaching high values from April 2015 to May 2016, and the Chl a concentration decreased and maintained at a lower level from May 2015 to July 2016 (Fig. 5).

4 DISCUSSION

Slopes of all beak variables in relation to ML for females were greater than those of males during 2013, 2014, and 2016, and slopes of UCL and LRL significantly differed between females and males in 2013 (P < 0.05). Meanwhile, K values of females were greater than those of males during 2013, 2014, and 2016, and K values significantly differed between females and males in 2013. Beaks are the important feeding organ of squids. The upper beak is a positive moving body, and the lower beak is the passive one to slice food coupled with upper beak while squid biting and then provide support for chewing (Raya and Hernández-González, 1998; Uyeno and Kier, 2007). The behavior and foraging habits of squids might be shifted with the morphological changes of beaks (Franco-Santos et al., 2014; Fang et al., 2017). As the beaks become more stiff and robust with the growth of beaks, the squids can forage on larger prey items (Castro and Hernández-García, 1995; Franco-Santos and Vidal, 2014; Hu et al., 2016a). Therefore, the somatic condition of squids is likely influenced by the beak shape, the somatic condition of females with greater slopes of beak variables was better than that of males.

We found that the somatic condition of females and males during the normal years (2013 and 2014) were significantly better than those from the El Niño year (2016), respectively (t-test, P < 0.01), while there was no significant difference between normal years (2013 and 2014) (Fig. 4). However, no significant difference was found for slopes of beak variables among years in this study. Therefore, the annual variability of somatic condition was not influenced by beak shape. In previous studies, the poor nutritional condition of D. gigas from the Gulf of California and Mexican waters during 1998 and 1999 was suggested to be caused by El Niño event 1997-1998 (Bazzino Ferreri, 2014). Pecl et al. (2004) suggested that females from cooler years had a better somatic condition and tended to have a higher level of reproductive investment. In this study, we found that the Chl a concentration in El Niño year was lower than that in normal years, which was also observed in previous studies (Barber et al., 1996; Radenac et al., 2012; Espinoza-Morriberón et al., 2017). Therefore, the common prey items of D. gigas were likely deficient, which was caused by the low primary production (Schwing, 2000; Robinson et al., 2013). Yu et al. (2016) studied the effect of the climate change on the habitat suitability of D. gigas and found that El Niño event weakened upwelling coupled with warm and nutrient-depleted waters, which were unfavorable for squid habitat and reduced catches. In this study, we suggested that the bad somatic condition of D. gigas during the El Niño year probably resulted from the low Chl a concentration in the seawater. It is adverse for D. gigas to grow and survive in the ambient environment in the El Niño year, and the abundance of D. gigas might decrease due to the unfavorable environment and the lack of prey items.

We figured out the CPUE during 2013, 2014, and 2016, and found that the CPUE of D. gigas in a normal year (5.22 t/d for 2013 and 4.29 t/d for 2014) was much greater than the El Niño year (2.61 t/d for 2016) (Fig. 6). In previous studies, the abundance of D. gigas in warm periods was also found less than that in normal years (Waluda and Rodhouse, 2006; Waluda et al., 2006; Yu et al., 2016).

In this study, the relationships between ML and beak variables were developed (Table 2). The squid beaks are hard to digest and usually found in the stomachs of marine predators (Lu and Ickeringill, 2002). This durability offers an opportunity to estimate squid size and biomass of their post-predation (Gröger et al., 2000; Lalas, 2009). Therefore, the beaks that remained in predators' stomachs can be used to estimate the body size of D. gigas based on the relationships developed in this study.

5 CONCLUSION

Having analyzed the D. gigas samples collected during 2013, 2014, and 2016, we revealed the variance of the somatic condition of D. gigas between normal years and the El Niño year. We believe that the relatively poor somatic condition of D. gigas during the El Niño year is resulted from the low Chl a concentration in the water. To make the Humboldt squid fishery sustainable and reduce the negative impacts of El Niño, more scientific management measures should be implemented in the year to come. The conclusion derived may need to be evaluated with more years of data in the future.

6 DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.

7 ACKNOWLEDGMENT

The data analysis and paper writing were done mainly when the senior author studied in Dr. CHEN Yong's lab in the School of Marine Sciences, the University of Maine, which is supported by Shanghai Ocean University and the University of Maine. We thank XING Lei (Ocean University of China) for statistical consultancy and advice on the figures.